The prognosis for patients with chronic myelogenous leukemia (CML) has improved only for patients who can receive marrow transplantation from a histocompatible sibling. The timing of the marrow transplant is made difficult by the high peritransplant mortality of 20% to 35% and a group of patients with a prolonged chronic phase of CML, which can be identified on the basis of prognostic indexes (age, percent blood myeloblasts, spleen size, and platelet count). We have developed a mathematic model and computer program that consider age, prognostic index, and projected survival rate by transplantation to balance the risk of peritransplant mortality against the risk of delaying the transplantation of patients with Philadelphia chromosome-positive CML. The computation assesses the risk of delaying transplantation; it does not offer the option of avoiding transplantation, since long-term survival ultimately requires transplantation. Three prognostic groups were considered as described by Sokal and co-workers (Blood 63:789, 1984) (I, best; II, intermediate; III, worst prognosis). The computation used the projected survival rates of transplantation from the Seattle experience and from the International Bone Marrow Transplant Registry. As an example of the model's utility, we have determined the ratio of the calculated life expectancy to the normal life expectancy for hypothetical patients up to 50 years of age in each of the three prognostic categories. A value of 20% is used for patients who successfully receive transplants after the onset of the accelerated phase. The analysis allows assessment of the risk of delaying transplantation for a finite time in patients with CML. The importance of the method rests in its consideration of multiple variables, including the peritransplant mortality, transplant projected survival before and upon entering the accelerated phase, age, prognostic group, and other risk factors. The program permits a change in these parameters as new information or advances in treatment occur. This analysis does not replace the diagnostic deliberations of the clinician. Rather, it provides a numeric framework for prognosis based on the currently available data. The physician in conjunction with the patient, not the algorithm, makes the decisions of whether and when to transplant.
Read full abstract